🐦 Soft Computing class project

Classification of bird species that appear on an image based on the use of a Convolutional Neural Network model (CNN)

Dataset features:

• Bird Species: 250;
• Training Images: 35215 (not balanced, however has at least 100 training image files per species);
• Validation Images: 1250 (5 per species);
• Test Images: 1250 (5 per species);
• Images Size: 224 x 224 x 3 color channels in jpg format;
• Species gender: 80% of total images are of male while the remaining 20% are of female

ML models developped:

• CNN + Search of optimal parameters using a Genetic algorithm + Data Augmentation techniques
• VGG16
• VGG19

🦜🦩🐤🦆🦉🐦🦃🦚🦅🦜🦩🐤🦆🦉🐦🦃🦚🦅